Curious about the stack. Are you using Tesseract, or did you build your own OCR?
Also, it might start to get REALLY expensive to run this once you start getting any appreciable number of queries.
I actually have nothing to do with the project. I just thought I'd share since I've never seen OCR that is free, easy, and doesn't even require registration.
Cool service, could easily get the text from the sidebar ad, but couldn't get the text from their own head logo.
Bookmarked.
I wish it supported Japanese. Combined with a GitHub repo to auto table scanned images would really help me out in my restaurant(I can't read Japanese very well, and we'll, in Japan all my receipts are in Japanese obviously haha).